design and construction of the billy bishop city airport pedestrian tunnel, toronto, canada

8/10/2019 Design and Construction of the Billy Bishop City Airport Pedestrian Tunnel, Toronto, Canada

1/8


2/8

behavior which occurs upon stress relief of the rock massand availability of fresh water.

3.2 Horizontal Bedding of Rock Mass

The short term kinematic stability of the backfilled TBM-driven secant bore arch was assessed on the basis of

numerical modeling using the discrete (distinct) elementprogram UDEC by Itasca Consulting Group, which offersthe advantage of modeling the entire constructionsequence of excavating and backfilling each individualdrift bore, followed by the main tunnel excavation andsupport as well as the potential to identify kinematic failuremechanisms. The UDEC analysis also considers stressand jointing anisotropy of the rock mass and models theprogressive stress re-distribution following eachconstruction stage.

UDEC uses a constitutive Mohr-Coulomb failurecriterion model which does not address the timedependent deformation (TDD) behavior of the GeorgianBay shale. However, potential kinematic failuremechanisms would be expected regardless of TDD

effects. Therefore, the use of UDEC was deemedappropriate to evaluate these kinematic failuremechanisms.

3.3 Time-dependent Deformation (TDD)

Hawlader, Lee, and Lo (2003) studied the impact ofapplied load on the swelling potential of different shalesamples. They concluded that the applied stress in oneprincipal stress direction reduces swelling strain not onlyin that direction but also in the perpendicular directions.Arup contracted the Itasca Consulting Group to implementthe numerical forumulations of Hawlader, Lee, and Lo(2003) into the Swello module in the FLAC 2D finitedifference progam to evaluate time-dependent swellingbehavior of the shale rock mass. The results of theseanalyses, and a comparison to measured values, arepresented later in the paper.

4 TUNNEL CONSTRUCTION DETAILS

4.1 Drift Bore Drilling and Temporary Support

The tunnel design concept proposed by TechnicoreUnderground involved the drilling and backfilling of seven1.85m diameter drift bores (1 to 7) within the crown of themain tunnel profile, two lower unfilled drift bores (8 and 9),and three main tunnel excavations center (cut 1),sidewalls (cut 2), and invert (cut 3) as shown in Figure1. The drift bores were sequentially drilled using two1.85m diameter tunnel boring machines, one of which isshown in Figure 2, constructed by TechnicoreUnderground. Each TBM was fitted with a total of 16300mm diameter disc cutters.

Each of the seven bores followed the same inclinationprofile of the main tunnel, namely a 1% incline from themainland portal to approximate mid-length, followed by a4% incline leading to the island portal. Lower bores 8 and9 were drilled at a constant inclination within the central

cut (no. 1) between main tunnel springline and invert leveland left open to provide both a void for the rockbreakerexcavating the central cut to break into and a temporaryventilation duct for dust control during excavation.

Temporary support within each drift bore consisted of1.2m x 2.4m sheets of 10mm thick plywood placedlengthwise along the crown of each drift bore and

expanded into place with circular steel ribs placed at 1.1mspacing immediately behind the tail shield of the TBM(Figure 3). This temporary support was provided toprevent pieces of shale from falling from the crown of thedrift bore. The center-to-center spacing of the fourprimary drift bores allowed for all to be left opensimultaneously prior to concrete backfilling, with thevertical load of the rock above being transmittedtemporarily through the rock pillars between adjacentprimary drift bores. The steel sets were typicallyrecovered during the retreat from each drift bore just priorto backfilling.

The structural compression integrity of the secant borearch was paramount to the stability of the main tunnelexcavation cavity, as it provided vertical support to the

crown rock load from above. During the excavation ofeach drift bore, some of the rock chips within thecutterhead were pulverized and subsequentlycompressed by the weight of the TBM shield as itadvanced against the wall of the bore at and belowspringline level into the equivalent of a stiff to hard clay ofapproximate 25mm thickness. As the presence of anysoft layers within the secant bore arch could compromiseits compressive stiffness and hence overall structuralintegrity, the rock abutment wall of each end primary bore(3 and 4) and the sidewalls of each of three secondarybores (5, 6, and 7) were cleaned of the aforementioneddeleterious materials prior to concrete backfillingoperations.

Upon completion of drilling each of the threesecondary crown drift bores (5, 6, and 7), the sidewalls ofthese bores were also inspected to ensure sufficientoverlap with the concrete backfill of the adjacent primarydrift bores. The design called for a minimum secantdimension of 300mm, which was governed primarily bythe minimum concrete shear area required to support theselfweight of the drift bore arch in the event that it was notfully engaged in compression. Despite a slight mis-alignment of the first drift bore, this overlap was achieved.

4.2 Drift Bore Backfilling

To avoid having to use the airport ferry to transportconcrete trucks to the island shaft, all drift bore concretebackfilling (with a 15 MPa compressive strength) wasadvanced from within the mainland shaft. Since thevertical position of each drift bore is 5m lower at themainland portal relative to its position at the island portal,drift bore backfilling operations had to be advancedthrough a 3m long concrete bulkhead cast within eachdrift bore at the mainland end. To prevent the buildup ofdrainage water behind the bulkhead prior to concreting, asmall drainage hole was provided within the bulkheadinvert, and the concreting tube was placed within thebulkhead crown, terminating just behind its interior end.


3/8

Figure 1. Tunnel profile drifts and main excavation

Figure 2. One of two 1.85m diameter TBMs for drift bores

Figure 3. 1.85m diameter drift bore temporary support

As the backfill concrete had to be pumped uphill toovercome a 5m elevation increase over a portal-to-portaldistance of 187m, it had to remain in a fluid state for theentire backfilling operation of each drift bore - potentiallyover an 8 hour duration for each 500 m3 volume.Technicore Undergrounds affiliate Tec-Mix developed aproprietary concrete mix for the drifts which allowed for

long duration fluid properties for pumping uphill, but didnot use traditional high cement content that would havemade the backfilled drifts difficult to mine through.

For the very first backfilling operation, a secondconcreting pipe was placed within the bulkhead crownwhich extended along the entire 1% grade and throughthe transition zone to the start of the 4% grade segment.This secondary pipe was installed in case the fluidconcrete could not be pushed along the entire 187mlength of the bore, either due to premature setting ordifficulties in overcoming the 5m elevation difference. Asit turned out, this second concreting pipe was notrequired, and it was subsequently deleted from allremaining six drift bore backfilling operations.

In an effort to ensure complete filling, a tube--

manchette (TAM) was installed within the crown of thefirst backfilled bore along its entire length to permitremedial grouting (if necessary), while truncated TAMtubes were used at the higher (island) end of allsubsequent bores. In fact, during the drilling of the centralsecondary bore, a tapered void (400mm maximum depthat the island portal) was encountered along the final 10mlength of each adjacent backfilled primary bore.Additional temporary support in the form of intermittentwooden wedge blocks, plywood, and shotcrete wasapplied within the sidewalls of the central bore toaccommodate a temporary span width more than doublethat of a single bore.

The cumulative duration of drift bore drilling andbackfilling took approximately 6 months, from December2012 to May 2013.

4.3 Main Tunnel Excavation

Central Cut 1 Excavation

Once drilling and backfilling of the series of seveninterlocking crown drift bores was completed (Figure 4),excavation of the main tunnel central cut (no. 1)commenced. This work was advanced from mainland toisland portal over a two month period using a Liebherr 934excavator with rockbreaker attachment (Figure 5). As ameans of dust control, a positive air stream was inducedtoward the island portal through one of the two open driftbores at/below the main tunnel springline with fansinstalled at both the mainland and island portals.The rock sidewalls were inspected by Arup staff on aregular basis, and the locations of vertical joints and waterseeps were noted. The primary vertical joint set(coincident with the major horizontal stress direction) wasoriented transverse to the tunnel axis. During theexcavation of the tunnel, vertical slices of rock typically fellfrom the tunnel face, so care was taken not to approachthe working face too closely.


4/8

Figure 4. Completion of secant drift bore backfilling

Figure 5. Main tunnel excavation with rockbreaker

Inspection of the underside of the crown drift boresindicated regular instances of transverse cracking, mostlikely resulting from shrinkage of the unreinforced backfillconcrete. These transverse shrinkage cracks were easilyidentified, as they were always accompanied by a nominalamount of water seepage. As the drift bore crown archwas intended to act solely in compression within thetransverse plane, the transverse cracks were not of greatconcern. Nevertheless, locations where a transversecrack intersected multiple bores were instrumented withadditional survey prisms and monitored for relativedisplacement across the crack. No such displacementwas observed.

Sidewall (Cut 2) and Invert (Cut 3) Excavation

After the central cut was complete, the main tunnel waswidened by the removal of the rock sidewalls, indicated ascut 2.

Cut 2 sidewall excavation was advanced primarily by areconditioned DOSCO roadheader (Figure 6) as well asan X-centric rock ripper, and took approximately 2 monthsto complete. The final sidewall excavation was achievedwith a small grinder attachment on an excavator arm. To

standardize the geometric profile of the main tunnelsidewalls, a template was fabricated and hung by a singlesliding, rotating pivot from temporary overhead railsaffixed to the underside of the crown drift bores (Figure 7).The template was rotated into place against the final (cut2) sidewall and checked for proper horizontal and verticalalignment. When the template was not in use, it was

rotated 90 degrees about its vertical axis, slidlongitudinally, and stored against the temporary cut 1sidewall.

Cut 2 was advanced in 3m increments. After eachincremental excavation, the final tunnel sidewallexcavation was bolted using an upper row of 3.5m longpolyester resin encapsulated steel dowels (32mm bardiameter) at 1.25m longitudinal spacing, followed by sixrows of 3.5m long Swellex MN24 bolts (installed on a1.5m staggered grid pattern). After cut 2 had advancedsome distance, the heads of the Swellex bolts were cutand a 50mm thick non-structural sealing layer ofpolypropylene fiber reinforced shotcrete was applied tothe exposed shale surface to prevent it from slakingdeterioration.

Invert (cut 3) excavation was advanced using abackhoe with mini-roadheader attachment (Figure 8), andtook approximately two weeks to complete.

4.4 Tunnel Waterproofing

The waterproofing system for the tunnel is comprised of a2.5mm thick PVC membrane which is separated from theshotcreted sidewalls by a layer of geotextile. Ribbedwaterbars were thermally welded to the PVC membranearound the entire tunnel profile perimeter at 12m intervals,coinciding with the construction joint locations of thepermanent tunnel arch lining. As the tunnel invert waspoured separately from the tunnel arch, longitudinalwaterbars were also installed along the entire tunnellength at the invert-arch interface.A gap grouting tube was installed within the crownalong the entire tunnel length, and remedial groutingtubes were installed at either side of the ribbed waterbarsand in the tunnel invert. It should be noted that the shaftwaterproofing was comprised of an HDPE membrane,hence a transition detail had to be developed between thePVC and HDPE membranes. This transition wasachieved using an intermediate Dilatec membrane andepoxy paste adhesive. In addition, a remedial groutingtube was also installed on the PVC-side of the transition,which can be injected with a polyurethane grout (ifnecessary).

4.5 Permanent Tunnel Lining

The permanent tunnel lining was constructed in twogeneral phases. First, the invert was cast in 18mmaximum lengths, followed by the tunnel arch, which wascast in 12m maximum lengths. The invert and arch poursare connected by a shear key. While there is continuity ofrebar over the invert construction joints, there is nocontinuity between the invert and arch pours or across thearch construction joints. As a means of secondary


5/8

Figure 6. DOSCO Roadheader

Figure 7. Tunnel sidewall (cut 2) profile template

waterproofing, a bentonite strip was installed at allpermanent tunnel lining construction joints.The invert is comprised of normal concrete, with a

minimum 28-day compressive strength of 35 MPa, whilethe arch is comprised of steel fiber reinforced concrete (30kg/m3 fiber dosage) with the same compressive strength.

The invert concrete pour was advanced efficientlyusing rebar mats and cages pre-fabricated by Technicoreaffiliate Ewing Fabricators, which were placed with a

Figure 8. Tunnel invert excavation

travelling crane with an extendable boom capable of liftingthe rebar mats and cages over a freshly poured section

and placing it on the invert waterproofing membranewithout damaging it. Invert construction, which tookapproximately three months to complete, would havetaken several weeks longer if the steel cages had beentied in the tunnel.

The relatively short tunnel length (187m) did not justifythe purchase of new formwork for the tunnel arch.Instead, the formwork used for the Devils Slide HighwayTunnels (San Mateo County, California, USA) waspurchased and modified by Technicore to suit the profileof this tunnel. The formwork ran along a pair of rails laidon the invert (Figure 11).

The tunnel invert and arch were poured to very tighttolerances, with variations of only a few millimetresrelative to theoretical, as measured by a Faro Focus 3DHigh Resolution LiDAR Scan.5 TUNNEL CONSTRUCTION MONITORING

5.1 Drift Bore Crown Stability during Main TunnelExcavation

As the central cut was advanced, a series of three opticalsurvey prisms were installed within the underside of thebackfilled crown drift bores, along with one in eachtemporary cut 1 sidewall, at 20m intervals along thetunnel. Regular monitoring of these survey prisms duringsidewall (cut 2) and invert (cut 3) excavations indicated astable rock mass during the works, as shown by samplesurvey results at Station 0+060m in Figure 13. Surveyresults were generally within 3mm of the initial baselinemeasurements, the variation attributed primarily to theprecision of the surveying method. Unfortunately, theinitial elastic movement of the drift bore arch crown wasnot captured during the initial excavation of central cut 1,although it is presumed to be significantly lower than thepredicted value.


6/8

Figure 9. Tunnel invert rebar placement

Figure 10. Tunnel invert concreting

Figure 11. Tunnel arch formwork and concreting

5.2 Main Tunnel Excavation Sidewall Convergence

Manual Convergence Tape

One method of measuring the horizontal convergence ofthe tunnel sidewalls during the main tunnel excavation

Figure 12. Completed permanent tunnel lining (June2014)

Figure 13. Optical survey data for crown and sidewalls

was with the use of a steel convergence tape. A hole wasdrilled into each of the optical survey prism points installedwithin the tunnel sidewalls (at approximate springlinelevel), which was then used to secure each end of theconvergence tape. The tape was manually tightened to astandard tension using a steel turnbuckle to obtain thehorizontal chord convergence reading.

The change in the horizontal chord dimension as afunction of time (days) after the central (cut 1) main tunnelexcavation is reported in Figure 14. With the exception ofthe station 0+060m readings, the horizontal chorddimension had reduced (converged) by only 1.5mm (or

less) up to 56 days after excavation. The single outlyingreading at 0+060m is believed to result from the west wallprism point not being fully fixed, as a consequence of avertical fracture oriented parallel to the tunnel axis behindthe prism point. It should be noted that as the initialconvergence readings were taken anywhere between 7and 22 days after the cut 1 excavation was advanced atany given tunnel station, the initial elastic sidewallconvergence was not captured in these convergencereadings.


7/8

Upon excavation of the tunnel sidewalls (cut 2), theoptical prisms installed within the cut 1 sidewalls weresequentially removed, reinstalled within the cut 2sidewalls, and rebaselined. Once again, the incrementalhorizontal chord convergence of cut 2 sidewalls atspringline level were recorded with time (days after cut 2excavation), as reported in Figure 15. With the exception

of the data recorded at station 0+020m, the incrementalshortening of the horizontal chord up to 84 days after cut2 sidewall excavation was 0.75mm or less.

While the manual convergence tape readings reportedin Figures 14 and 15 demonstrate a general trend in theshortening of the horizontal chord with time, the data arealso characterized by a significant degree of day-to-dayvariability, which is attributed to the combined effect ofambient temperature fluctuations within the tunnel, alongwith standard operational errors. The relative effects ofthese error sources on the convergence measurementsare more pronounced because of the very small absolutevalues which were recorded.

MPBXs

In an effort to obtain more reliable tunnel sidewallconvergence measurements, two multi-point boreholeextensometers (MPBXs) were installed at station 0+020m(two tunnel diameters from the mainland portal), one ineach sidewall, in mid-July of 2013. The MPBXs wereinstalled within 3m long alcoves excavated from thecentral cut 1 sidewall to the final main tunnel sidewall (cut2) profile.

The MPBX data for the eastern wall has previouslybeen reported by Hurt, et al. (2014), and is updated inFigure 16. The overall MPBX was 12m in length, withvalues of inward deformation (convergence) measured atthe tunnel sidewall and at 2m, 5m, and 8m behind thewall, all measured relative to a presumed point of fixity12m behind the wall. Very little wall movement(~0.05mm) was recorded until the east sidewall cut 2excavation work was completed both up-station anddown-station from the 3m long alcove (centered at station0+020m) between August 29th and September 6 th, 2013,which resulted in approximately 0.1mm of immediateelastic sidewall movement followed by a nominal degreeof time-dependent movement. A second discrete jump inthe tunnel sidewall movement (~0.2mm) was observedwhen the invert (cut 3) was excavated between October18-19, 2013, which was again followed by time-dependentmovement. The tunnel sidewall was pushed back slightlyduring the concreting of the permanent tunnel arch in mid-March, 2014, but recovered rather quickly;the MPBXpoints within the rock mass have since experienced anominal residual effect.

After nearly one year of MPBX measurements, thetotal inward movement of the eastern wall was only0.6mm, with about half of this movement being attributedto elastic movements induced by the sidewall (cut 2) andinvert (cut 3) excavations. The remaining time-dependentinward movement of the eastern wall (after completion ofthe cut 2 sidewall excavation) of approximately 0.25mm isgenerally consistent with the typical post-cut2 shortening

Figure 14. Horizontal convergence tape measurementsafter central (cut 1) tunnel excavation

Figure 15. Horizontal convergence tape measurementsafter tunnel sidewall (cut 2) excavation

Figure 16. Updated eastern wall MPBX readings


8/8

Figure 17. Time-dependent tunnel sidewall swelldeformations - MPBX readings compared to predictedvalues (Hurt, et. al., 2014)

of the horizontal chord of less than 0.75mm as indicatedby the manual convergence tape readings (Figure 15).

Figure 17 shows the time-dependent horizontal tunnelsidewall swell deformations (i.e., total deformation minuselastic deformation) compared to a FLAC 2D Swellomodule prediction using the numerical analysis procedureof Hawlader, Lee, and Lo (2003). It should be noted thatthe tunnel sidewall swelling prediction presented in Figure17 was made with the aid of a back-analysis of time-dependent horizontal free swell measurements obtainedfrom borehole inclinometers installed in the rock walls ofthe mainland shaft, which suggested a horizontal freeswell potential of between 0.3% to 0.4% (per log cycletime), an in-situ horizontal stress of approximately 5 MPa,and a critical (swell suppression stress) of about 3 MPa.The data in Figure 17 show generally good agreement,with the measured time-dependent MPBX convergencebeing somewhat lower than that predicted from thenumerical model. One reason for this is that theexcavated rock walls inside the tunnel were relatively dry,as there was very little induced damage to the rock massfrom the roadheader excavation (relative to a drill-and-blast excavation, for instance), although the mainlandshaft sidewalls on which the back-analyzed inputparameters were based did have adequate access tofreshwater to facilitate swelling.

Figure 18. Comparison of tunnel sidewall swelldeformation predictions closed form solution versusnumerical analysis (Hurt, et. al., 2014)

The numerical procedure of Hawlader, Lee, and Lo(2003), which was implemented by Arup using the FLAC2D Swello module, should be viewed as a majorimprovement over the older closed-form solution (Lo andYuen, 1981), as it considers the non-linear time- andstress-dependency of the key swelling parameters,typically resulting in less conservative (but more realistic)

predictions of tunnel sidewall convergence deformation.Figure 18 shows the large discrepancy in the closed-formprediction compared to that of the numerical analysis.After a period of 100 days, the closed-form method hadpredicted sidewall convergence values of between 40 and140mm (a function of the range of input parameters),which was far in excess of the refined numericalprediction (and MPBX observation) of approximately0.25mm after a similar length of time.

6 SUMMARY

This paper has presented several design and constructionaspects associated with the Billy Bishop City AirportPedestrian Tunnel in Toronto, Canada. The tunnel was

constructed in shale bedrock using a unique system ofpre-support consisting of seven interlocking secant driftbores drilled by tunnel boring machines and backfilledwith mass concrete. The construction monitoring resultsdemonstrated that the main tunnel cavity excavatedunderneath the tunnel arch pre-support was stable. Whilethe tunnel convergence measurements were far less thanthose predicted by a closed form solution for rock swell,they agreed reasonably well with a non-linear time-stepnumerical procedure, with input parameters calibratedwith the observed time-dependent vertical shaft wallmovements.

ACKNOWLEDGEMENTS

The writers would like to acknowledge the contribution ofa number of individuals to the paper, including SethPollak, Sean Lee, and Amirreza Ghasemi (Arup), as wellas Gary Benner, David Marsland, and Joey DiMillo(Technicore).

REFERENCES

Hawlader, B.C, Lee, Y.N., and Lo, K.Y. 2003. Three-Dimensional Stress Effects on Time-DependentSwelling Behavior of Shaly Rock, CanadianGeotechnical Journal , Vol. 40, pp. 501-511.

Hurt, J., Lee, S., Ghasemi, A., Pollak, S., and Cushing, A.2014. Time Dependent Movements on the Billy BishopToronto City Airport Pedestrian Tunnel, Ontario,Canada. North American Tunneling Conference , LosAngeles, CA, USA.

Lo, K.Y. and Yuen, C.M.K. 1981. Design of Tunnel Liningin Rock for Long Term Time Effects. CanadianGeotechnical Journal , Vol. 18, pp. 24-39.

design and construction of the billy bishop city airport pedestrian tunnel, toronto, canada

Documents